Autonomous exploration and semantic mapping of a mobile robot using efficient frontier selection

The frontier-based exploration method is a common approach used to navigate unknown region. Although existing frontier-based exploration is computationally simple in the exploration of the frontiers, this algorithm can cause unnecessary motion, thereby increasing the exploration time and decreasing the exploration success rate. To resolve this issue, we propose a method to improve the performance of the existing frontier-based exploration method by using an A-star path planner and a frontier selection algorithm that consider the direction of a mobile robot. In addition, we employed a real-time object detection technique to build a semantic map while exploring an unknown region. To verify our proposed approach, we performed experiments to compare the existing approach with our proposed method in the clutter environment(in real and Gazebo simulation). Through the experimental results showed that our method shortened the exploration path and time. © ICROS 2020.